Archive for the 'Consulting' Category


Oracle Mix

Seems ThoughtWorks is up to it again, this time with Oracle: they’ve made a social networking application for Oracle’s customers, using jRuby and Rails.

I’m overjoyed that these enterprise applications are emerging, with big company support.  I have been “agitating” the idea of Ruby and Rails in the enterprise for the last eighteen months but have always met resistance in the Australian market.  The larger companies here are very risk averse and haven’t been able to grasp the value of a rapid development environment with strong test driven development underpinnings, but as these forward thinking organisations such as ThoughtWorks and Oracle demonstrate, there are great opportunities to be had.

Can you believe they took only 5 weeks to build this app!  I can, but then again I use Rails every day.

EJA Forums Sydney – October 19th 2006

On Thursday 19th October I visited the Sydney edition of the Enterprise Java forum at the MLC Center. From what I understand EJA forums have been running in Melbourne for longer, and there’s an event down there monthly, whereas they are just establishing themselves in Sydney, and the events are only quarterly. The Melbourne program was slightly more extensive, but Sydney’s program was pretty busy too, with six speakers spread over half a day.

Firstly, my notes on the speakers:

Jim Webber, Geurilla SOA — was engaging and humourous, made a good case for his expertise and expounded a sensible rationale for adopting SOA. The main thrust of his argument was that Web Services alone are robust enough to provide the benefits of SOA, without having to rely on vendor solutions and the lock-in they entail. A useful analogy for this proposition was comparing the telecommunications network to the internet, where the internet has won because of its ignorance and simplicity, compared to the complex routing and domain knowledge of the telecomms networks.

Peter Evans-Greenwood, The Boundaryless Organisation — Peter’s speech was largely strategic, and I suspect was over the heads of most of the mainly technical attendees. Certainly from a purely Java perspective there was little of value. Probably it was just the wrong forum for this type of talk.

Ben Alex, Next-Gen POJO Architecture — Ben’s points were largely that while architecture has improved in recent years, it could be better with some simple tweaks that can vastly improve the success of new systems. Firstly he wants us to get the terminology right, some of the patterns being used commonly are being incorrectly labelled. More importantly, the idea of “domain driven design”, with a better separation of concerns and a longer shelf life was promoted. He referenced the book “Domain Driven Design” several times, and its been on my reading list for a while, so I should get to it soon.

Paul Scott-Murphy, Architecting Event Driven Systems — I’d mostly seen this talk before (at a TIBCO event), but with some future projects in mind, this was particularly interesting. Paul demonstrated TIBCO products and how they can help create an event driven system, using Confluence and the events it generates as an example.

Cristina Cifuentes, Java on Wireless Sensor Devices — interesting from a hobby technology point of view.

Mike Cannon-Brookes, Pragmatic Clustering — Mike spoke about Tangosol, and lessons learned from Atlassian’s efforts in clustering Confluence. Most of all it was interesting to see his thought process in reaching the decision to use Tangosol… they started off not knowing much about it, and so we got to learn from the mistakes they made along the way. I wish I’d taken better notes during this (update: the presentation is here).

As for the rest of the event. The networking opportunities seemed limited at first glance. There were a couple of consulting companies, and the bulk of the audience was from the Reserve Bank of Australia. But aside from these there wasn’t much on offer. As with all technical talks the crowd was a little shy and geeky, but ready to chat easily enough. Also, the food was great!

GLiNTECH in BRW Fast 100

The company that I work for, GLiNTECH, was placed at 19th in BRW‘s 2006 Fast 100 list.  I managed to win the office pool by having a guess closest to the right answer.
Anyway, congratulations guys, lets get in the top 10 next year!

Tufte on Powerpoint

Edward R. Tufte – The Cognitive Style of PowerPoint

Yes, I realise the irony of using bullet points to summarise this too. But some notes are better than no notes.

  • Presenters love PowerPoint, not audiences
  • PP is so low on information density that truths must be abbreviated to fit
  • Bullets can’t show complex relationships, can only show ONE of these:
  • Sequence
  • Priority
  • Or simple set membership
  • Tufte recommends sentences with subjects and verbs rather than bullets
  • Quote from Richard Feynman (re: NASSA and his role on commission investigating Challenger disaster in ’86):
  • Then we learned about “bullets” – little black circles in front of phrases that were supposed to summarise things. There was one after another of these goddamn bullets in our briefing books on slides

    • Thin content -> boring presentations -> adding PPPhluff
    • Greater detail means greater clarity and gives context
    • Branding on PP slides is insidious
    • “The way to make big improvements in a presentation is to get better content”
    • Avoid hierarchies of bullets
    • Never read from slides
    • Never use PP for paper reports
    • Instead of PP, use handouts of additional material that is dense in information

    TechSession: Ruby and Ruby on Rails

    I delivered an introductory TechSession for GLiNTECH on Ruby and Ruby on Rails a couple of weeks ago. I tried some different presentation techniques to the usual (inspired by Presentation Zen) and with some subject matter that lent itself to the task, came up with something that in my opinion did a pretty good job of holding the audience’s attention. I also got to use the remote control (via SofaControl) of my newish MacBook Pro, which made me feel like a massive nerd.
    Unfortunately the style of presentation means that the slides are pretty much useless by themselves, and the session was not recorded. By way of summary, a very brief summary of my main points:

    Ruby

    • Simple, elegant and OO
    • Principle of least surprise
    • Portable
    • More powerful than “scripting language” implies
    • Language features: Mixins, Blocks, Closures, Continuations, Aliasing
    • Issues: performance & vendor support
    • Advantages: “less” – lines of code, configuration, bugs, frameworks
    • Can give you new perspectives on your main commercially employed language

    Ruby on Rails

    • MVC based web framework
    • Convention over configuration
    • Quick to get started, but completely extensible and controllable
    • Features: Migrations, ActiveRecord, Scaffolding
    • Has great marketing — and marketing matters (see Sun/Java, Microsoft/C#)
    • WILL influence your web development, even if you never use RoR (see Grails, Trails etc)
    • Advantages: fast start, rapid feedback, less code

    Some of the audience hung around and we spoke about technical details as well as strategy and marketing angles. An interesting topic and some great questions made presenting this rather enjoyable.